estimate
Can I run Qwen3.8-27B on an RX 7900 XTX?
Yes, fully in VRAM (estimate)
Can I run Qwen3.8-27B on an RX 7900 XTX? Yes, fully in VRAM (estimate). About 17.8 GB at Q4_K_M (estimate).
VRAM per quant for Qwen3.8-27B
Updated . The later of the hardware price date and the formula date.
| Quant | Context | Weights | KV | Overhead | Total | Usable | Fit | Est. tok/s |
|---|---|---|---|---|---|---|---|---|
| Q4 Q4_K_M | 8,192 tokens | 16.5 GB measured GGUF file size | 0.5 GB estimate | 0.8 GB estimate | 17.8 GB estimate | 23.4 GB estimate | full | 22 tok/s estimate |
| Q5 Q5 (~0.71 bytes/param) | 8,192 tokens | 19.7 GB estimate | 0.5 GB estimate | 0.8 GB estimate | 21.1 GB estimate | 23.4 GB estimate | full | 18.9 tok/s estimate |
| Q8 Q8 | 8,192 tokens | 29.7 GB estimate | 0.5 GB estimate | 0.8 GB estimate | 31.1 GB estimate | 23.4 GB estimate | no | — |
| FP16 FP16 | 8,192 tokens | 55.6 GB estimate | 0.5 GB estimate | 0.8 GB estimate | 56.9 GB estimate | 23.4 GB estimate | no | — |
| Q4, longer context | 32,768 tokens | 16.5 GB measured GGUF file size | 2.1 GB estimate | 0.8 GB estimate | 19.4 GB estimate | 23.4 GB estimate | full | 21 tok/s estimate |
Formula estimate, not a benchmark. Method →
community measurement, not ours
Canonical single-GPU figures on that results page: mean of five runs, one RTX 3090, the other replica stopped, the ~1K-token prompt column, 512-token greedy completions, reasoning disabled, 250 W power cap. The page says every rung is the same 4-bit checkpoint. llama.cpp out of the box 27.6 tokens per second; llama.cpp tuned (FA + q8 KV + MTP) 42.6; stock vLLM + MTP k=4 85.6; patched vLLM + MTP k=4 167.8; patched vLLM + DFlash2 k=7 212.1. The page dates the single-replica correction 31 August. results.html was last committed on 18 September 2026, which is the published date stored here. Not a localLLMrig measurement. Calculator estimates keep benchmarkSourceUrl null.
| Hardware | Tokens per second |
|---|---|
| RTX 3090, llama.cpp out of the box | 27.6 tok/s community measurement, not ours |
| RTX 3090, llama.cpp tuned (FA + q8 KV + MTP) | 42.6 tok/s community measurement, not ours |
| RTX 3090, stock vLLM + MTP k=4 | 85.6 tok/s community measurement, not ours |
| RTX 3090, patched vLLM + MTP k=4 | 167.8 tok/s community measurement, not ours |
| RTX 3090, patched vLLM + DFlash2 k=7 | 212.1 tok/s community measurement, not ours |
Qwen3.8-27B on RTX 3090, published 2026-09-18
This measurement stays separate from the formula estimate above.
Cheapest catalog option that fits at Q4
GeForce RTX 3090 24GB, used (median of sold listings), €1,224, dated 2026-10-16. No buy link.
This hardware
$1,404€1,254
used (3-month median)
price as of 17 Oct 2026
new in stock, €3,972 at GPU PRIX on 17 Oct 2026 (unusually high): $4,447€3,972. GPU PRIX. More than 1.5× the €1,254 3-month median, so this listing is not the headline.
GPU PRIXChecked GPU PRIX on 17 Oct 2026. The page says last updated Oct 17, 2026. The 3-month median is €1,254. The only in-stock board, ASUS TUF Gaming RX 7900 XTX at €3,972, is more than 1.5× that median, so the median is the stored price. The header current low of €43 is a DisplayPort cable, not the card. XFX Speedster was €1,049 and out of stock. MSI Gaming Trio €3,047 and POWERCOLOR Red Devil €3,152 were out of stock. On 16 Oct 2026 the stored new figure was the out-of-stock price, about €1,050. 12-month median €1,101. Not a pcprice.watch sold-listing median.
Embed this
Qwen3.8-27B on RX 7900 XTX
[](https://localllmrig.com/can-i-run/qwen3.8-27b/rx-7900-xtx?utm_source=badge)
Questions
- Can I run Qwen3.8-27B on an RX 7900 XTX?
- Yes, fully in VRAM (estimate).
- How much memory does Qwen3.8-27B need at Q4?
- About 17.8 GB at 8,192 tokens (estimate): 16.5 GB weights (measured GGUF file size), 0.5 GB KV cache, 0.8 GB overhead.
- How fast is Qwen3.8-27B on RX 7900 XTX (estimate)?
- About 22 tok/s at Q4 (estimate, not a benchmark).
Rig Report waitlist
A written hardware plan for one model and budget. Planned price, not for sale yet.
Sign up with consent and download GPU Buying Cheat Sheet (EU, Oct 2026). The file is not emailed.
Read next: Building a local LLM rig in the RAMageddon.